- Email: [email protected]
Transient Expression of Type III Collagen by Odontoblasts: Developmental Changes in the Distribution of Pro-α1 (III) and Pro-α1 (I) Collagen mRNAs in Dental Tissues
Matrix Vol. 13/1993, pp. 503-515 © 1993 by GustavFischer Verlag, Stuttgart· Jena . New York
Transient Expression of Type III Collagen by Odontoblasts: Developmental Changes in the Distribution of Pro-u1 (III) and Pro-u1 (I) Collagen mRNAs in Dental Tissues PIRJO-LiISA LUKINMAA1, ANNE VAAHTOKARI 2 , SEPPO VAINI0 2 , MINNA SANDBERG 3 , JANNA WALTIM02 and IRMA THESLEFF 2 1 2
3
Department of Oral Pathology, Institute of Dentistry , University of Helsinki, Helsinki, Department of Pedodontics and Orthodontics, Institute of Dentistry , University of Helsinki, Helsinki and Department of Medical Biochemistry, University of Turku, Turku, Finland.
Abstract
The expression of Pro-al (III) and pro-al (I) collagen mRNAs in mouse and human dental tissues during tooth development and after its completion was analyzed by in situ hybridization, with use of [35 S]-labeled RNA probes. The expression of pro-a I (III) mRNA was also compared to that of the protein product, as localized by immunostaining with polyclonal antibodies to type III collagen and the N-terminal propeptide of type III procollagen. Contrary to many previous reports, our results suggest that odontoblasts express type III collagen. While proal (III) transcripts were less intensely expressed in odontoblasts than pro-a I (I) transcripts, the amounts of both mRNAs increased in odontoblasts with progressing dentin formation, and decreased toward its completion. In contrast to Pro-al (III) mRNA, pro-al (I) mRNA was still detectable in odontoblasts of fully developed teeth. Type III collagen immunoreactivity was observed in the early predentin, and again in predentin toward the completion of dentinogenesis, when mRNA was no longer detected. Also in the pulp, the protein product, unlike pro-al(III) mRNA, was relatively strongly expressed. Hence, these immunostaining patterns were inversely related to the expression of Pro-a1 (III) mRNA, suggesting accumulation of the protein. The mesenchymal cells, when condensed in the region of the future mandibular bone, expressed proal (III) mRNA intensely, whereas osteoblasts expressed pro-aj (I) but not pro-aj (III) transcripts strongly. Cell type- and developmental stage-related differences in the expression of the two mRNAs suggest that type Iltype III collagen ratio influences the structure of dental tissues. Key words: in situ hybridization, immunohistochemistry, pro-a I (III) and Pro-al (I) collagen mRNAs, (pN) type III collagen, teeth. Introduction
The main structural component of the organic matrices of predentin and dentin of various species is type I collagen (for reviews, see Butler, 1984; Linde, 1989). While type III collagen remains a major constituent of the dental pulp even after the completion of tooth development (Shuttleworth et al., 1978; van Amerongen et al., 1983; Linde,
1985; Tsuzaki et al., 1990), it is generally thought that odontoblasts do not express type III collagen at the mRNA (Andujar et al., 1991) or protein level (Lesot et al., 1981; Ruch, 1987; Andujar et al., 1988). Although early biochemical studies have suggested the presence of type III collagen in dentin (Volpin and Veis, 1973), in a number of later biochemical (Scott and Veis, 1976; Munksgaard et al., 1978; Dodd and Carmichael, 1979; Munksgaard and Moe,
504
P.-L. Lukinmaa et al.
1980; Sodek and Mandell, 1982; Karjalainen and Soderling, 1984) and immunohistochemical (Cournil et ai., 1979; Wright and Leblond, 1980; Takita et ai., 1987) studies on teeth of different species it has not been possible to demonstrate that odontoblasts synthesize type III collagen or that it is present in predentin. On the other hand, some immunohistochemical studies suggest that terminal differentiation of odontoblasts might not be accompanied by the cessation of type III collagen production. For example, type III collagen immunoreactivity has been detected in odontoblasts of fetal porcine and adult human teeth (Tung et ai., 1985; Karjalainen et ai., 1986). Also the first formed predentin of developing mouse teeth (Thesleff, et ai., 1979, 1991), dentin of young mice (Nagata et ai., 1992) as well as predentin of fully developed human teeth (Becker et ai., 1986; Karjalainen et ai., 1986) have been found to immunostain for type III (pro )collagen. Osteoblasts are believed to represent another major cell lineage that does not express type III collagen, as demonstrated even at the mRNA level (Sandberg and Vuorio, 1987; Sandberg et ai., 1988; Andujar et ai., 1991). However, recent immunoelectron microscopic data show that not only developing but also adult human cortical bone contains type III collagen specifically retained in discrete fiber bundles (Keene et ai., 1991). While the genes coding for the pro-a chains of types I and III collagen are differentially expressed in distinct tissues and at various stages of development (Burgeson, 1988; Sandberg et ai., 1989; Andujar et ai., 1991), the production of these proteins usually correlates well with the expression of the corresponding mRNAs (Sandberg and Vuorio, 1987; Sandberg et ai., 1989). With regard to teeth, once tooth development has been completed, the rate of dentin deposition is markedly decreased and unlike bone, dentin is not remodeled. Furthermore, as the half-life of collagen is remarkably long, the presence of the protein product in predentin does not necessarily imply simultaneous transcription in odontoblasts. We have previously shown that after the completion of human tooth development, odontoblasts continue to express mRNA for pro-a2(I) collagen chain (Lukinmaa et ai., 1992). The aim of this study was to analyze the expression of the genes encoding the pro-al chains of types I and III collagen in dental tissues. For this purpose, the expression of the two mRNAs was determined in mouse and human dental tissues during tooth development and after its completion, with radiolabeled RNA probes. The expression of pro-al (III) collagen transcripts was also compared to that of the protein product, as localized by immunohistochemical staining with antibodies to type III collagen and the N-terminal propeptide of type III pro collagen.
Materials and Methods Mouse Teeth
The regions of the mandibular molar tooth germs were dissected from 13-day-old embryonic (vaginal plug = day 0), newborn, and 4-day-old postnatal mice (CBAxC57BL). As determined morphologically, the developmental stages of the first and second molars covered the bud stage (not analyzed), the early (odontoblasts not polarized) and late (deposition of predentin initiated in the cuspal region) bell stages, and a stage where dentin was being mineralized throughout the tooth crown. Human Teeth
The germ of the mandibular left second premolar tooth (mineralization of dentin started at the tip of the main cusp) was accidentally removed from a four and a half-year-old girl, in association with the therapeutic removal of the deciduous second molar. A neonatal mandibular incisor (mineralization of dentin in progress throughout the tooth crown) was obtained from a newborn baby. Two mandibular third molars, in which coronal dentin formation was advanced but root development had not begun, were removed from a 14-year-old girl. One developing (about half the root lengths completed) and several fully developed third molars, one of them (to be analyzed by in situ hybridization) affected by occlusal caries, were from young adults. These teeth were also removed for therapeutic reasons. Preparation of the tissues for in situ hybridization
The mouse molar regions were fixed overnight with 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS) at 4°C, dehydrated through a rising ethanol series and xylene, embedded in paraffin, sectioned at 71-lm thickness on silane-coated slides, and air-dried at 37°C. The human teeth with the attached soft tissues were fixed with 10% neutral buffered formalin or 4% PFA for 2-5 days. To facilitate the penetration of the fixative to the pulp chamber of the fully developed tooth, about one third of the root lengths was dissected away immediately after extraction. The teeth were demineralized with ethylenediaminetetra-acetic acid (0.33 MIl) for 4-12 weeks (depending on the stage of development) at 4°C, rinsed overnight with running tap water, and processed into paraffin sections as described above. The slides were stored in tight boxes at 4 °C until used. For a detailed histological examination, some sections, corresponding to those used in the hybridization experiments, were stained with hematoxylin and eosin.
Pro-a}(III) and Pro-al(I) Collagen mRNAs in Teeth
Preparation of the probes A 372-bp Pst I-Pvu II fragment from clone pHCALl, specific for human pro-a} (I) collagen mRNA (Vuorio et ai., 1987), was sub cloned into pGEMl plasmid (Promega Biotec, Madison, WI). The plasmid was linearized with Pst I or EcoR I restriction endonuclease (Boehringer Mannheim, Germany), and eSS]-UTP-labeled single-stranded antisense and sense pro-a} (I) RNAs were synthesized in vitro (Wilkinson and Green, 1990), with T7 or SP6 RNA polymerase (Promega Biotec), respectively. A 295-bp Pst I-Pst I fragment from clone pHFS3, specific for human pro-a} (III) collagen mRNA (Sandberg et al., 1989), was also subcloned into pGEM1 plasmid, which was linearized with Hind III or
505
EcoR I restriction endonuclease (Boehringer Mannheim). For synthesis of the antisense and sense pro-a} (III) RNAs, T7 or SP6 RNA polymerase was used. RNAs were shortened to an average length of 100 bases by limited alkaline hydrolysis. For isolation ofprobes of the desired length, gelfiltration chromatography was performed on a Sephadex G-50 (Pharmacia, Uppsala, Sweden) column. After precipitation with ethanol, the probes were dissolved in the hybridization buffer containing 60% formamide, and diluted to about 5.5 X 104 dpm[ll-1. Homologies between the nucleotide sequences of the pro-a} (I) collagen probe and the corresponding regions of pro-a2(I) and pro-a}(III) mRNAs were 64% and 59%,
Fig. I. Expression of pro-al(III) collagen mRNA (A-D) and the N-terminal propeptide of type III procollagen (E-G) in transverse sections through a 13-day-old mouse embryo mandible. In situ hybridization patterns with the antisense (A, B) and sense (negative control) (C, D) probes, and bright-field (A, C) and dark-field (B, D) images are shown. The level of pro-adIII) mRNA expression is moderate in the mesenchyme (m). Those mesenchymal cells that are condensed in the region of the future mandibular bone (arrows) express pro-al(III) transcripts relatively strongly. A weak signal is also detected in Meckel's cartilage (c). (E) Immunostaining for the Nterminal propeptide of type III pro collagen is moderate in the mesenchyme (m), whereas Meckel's cartilage (c) is negative, unless the section is pretreated with hyaluronidase, which facilitates reactivity especially in its peripheral zone, as seen in (F). (G) No staining is seen in a hyaluronidase-pretreated control section incubated with normal rabbit immunoglobulin. Bar is 50 ftm.
506
P.-L. Lukinmaa et al.
respectively. The nucleotide sequence of the pro-at (III) probe showed only 51 % identity with the corresponding regions of the human pro-at (I) and pro-a2(I) collagen mRNAs. The probes did not contain the sequence of the 87
nucleotides in the middle of the carboxy-terminal propeptide domain, which is virtually identical in mRNAs for fibril-forming collagens (Sandberg and Vuorio, 1987).
Fig. 2. Pro-at (I) (A, B, E) and pro-at (Ill) (C, D, F) collagen mRNA expression in sagittal sections through the mandibular molar region of a newborn mouse. Bright-field (A, C, E, F) and dark-field (B, D) images of in situ hybridization patterns with the antisense probes are shown. The first (right; A-D) and second (left) molar tooth germs correspond to the late and early bell stages, respectively. Symbols: 0, odontoblasts; p, dental pulp/papilla (A-D); f, dental follicle (A-E); a, ameloblasts; sr, stellate reticulum; m, oral mesenchyme; b, mandibular bone. (A, B) Expression of pro-at (I) mRNA in the cuspal area (arrows) of the second molar germ is slightly stronger than elsewhere in the dental papilla. Polarized odontoblasts (first molar) express pro-at(l) mRNA very strongly. Hybridization is relatively weak in the dental pulp, follicular tissue and oral mesenchyme, whereas the signal intensity in the developing mandibular bone is fairly high. Ameloblasts and stellate reticulum are negative. (C, D) Virtually no expression of pro-at (III) mRNA is seen in the dental papilla of the second molar. With deposition of predentin (first molar), the level of expression increases in odontoblasts, and the signal in the pulp is also more intense than in the dental papilla of the second molar. Grain condensations (arrowheads) are seen around the epithelial cervical loop tips. Hybridization is moderate in the follicular tissue and relatively strong in the oral mesenchyme. Ameloblasts and stellate reticulum are negative. Higher magnifications (E, F) show that osteoblasts (arrows), especially those on the oral aspect of the developing mandibular bone, express (E) pro-at(I) but not (F) pro-at (III) collagen transcripts strongly. Expression of pro-at (III) mRNA in fibroblasts between the bony trabeculae is intense. Also note that the signal intensity for pro-at (III) transcripts is lower in odontoblasts of the first molar cusp tip and higher in the oral mesenchyme than that for pro-at (I) mRNA. Oral epithelium (e) is negative (F). Bar is 150 ~m in A-D and 50 ~m in E and F.
Pro-at (III) and Pro-at (I) Collagen mRNAs in Teeth
In situ hybridization In situ hybridization was performed as described by Wilkinson and Green (1990). Tissue sections were rehydrated, treated with proteinase K and acetylated. Before hybridization, the probes were heated at 80 DC for 2 min. The sections were hybridized overnight at 50-52 DC, in a tight box moistured with the hybridization buffer containing 60% formamide. The hybridized sections were washed under highly stringent conditions, including a wash at 65 DC (30 min), a treatment with ribonuclease A (Boehringer Mannheim, Germany), and another 30 min wash at 65 DC. The slides were air-dried, dipped in autoradiographic emulsion (Kodak NTB-2; dilution 1:1), again dried in the presence of silica gel, and exposed at 4 DC for 10-15 days. After development and fixation, the sections were shortly counterstained with hematoxylin, and mounted with DePex (BDH, Poole, England). The slides were examined under a Leitz Orthoplan microscope and selectively photographed using Ilford Pan F film.
507
Preparation of the specimens for immunohistochemistry Paraffin sections of mouse and human tissues were prepared as described above. Mouse tissues were also fixed with 94% cold ethanol and human teeth with ethanol at about 22 DC. To enhance the reactivities, some rehydrated sections were pretreated with pepsin (Merck, Darmstadt, Germany; 0.4% in 0.1 M HCI, 37"C, 40 min) and/or ovine testicular hyaluronidase (Boehringer Mannheim, Germany; 3500 and 7000U/ml PBS, pH6.0, 37"C, 30 min). Some human teeth were frozen with liquid nitrogen and processed into sections as earlier described (Lukinmaa and Waltimo, 1992).
Antibodies Specific rabbit anticalf skin type III collagen antisera, cross-reactive with tissues of other species (Peltonen et al., 1982), were generously donated by Dr. Leena Peltonen (National Public Health Institute, Helsinki, Finland). Immunoaffinity-purified rabbit antibodies specific for the N-terminal propeptide of human type III pro collagen
Fig. 3. (A) Immunostaining for type III collagen in the late bell stage first molar tooth germ of a newborn mouse is moderate in the pulp (p) and cytoplasmic reactivity is also seen in secretory odontoblasts (0). Predentin (pd) in the enzymatically untreated section is virtually negative. The dental follicle (f) stains very intensely, whereas the ameloblastic layer (a) is negative. (B) Control section treated with normal rabbit serum is negative. (C) Antibodies to the N-terminal propeptide of type III procollagen stain the pulp (p) of the first molar moderately, whereas the odontoblastic layer (0) is negative. The strong reactivity in the dental basement membrane zone (arrows) disappears toward predentin (arrowheads) in the cuspal area, where (D) hyaluronidase pretreatment facilitates weak staining in the innermost predentin (arrowheads) and also intensifies reactivity in the pulp (p) and basement membrane. Staining in the dental follicle (f) is intense, whereas the ameloblastic layer (a) and stellate reticulum (sr) are negative. Note the strong reactivity around the epithelial cervical loop (cl) (C, D). Bar is 30 ~m in A and B, and 100 ~m in C and D.
508
P.-L. Lukinmaa et al.
(Niemela et aI., 1985) were a kind gift from Drs. Leila and Juha Risteli (Department of Medical Biochemistry, University of Oulu, Oulu, Finland). Immunohistochemical staining
Frozen and paraffin sections were stained with the immunoperoxidase method as earlier described (Lukinmaa and Waltimo, 1992), according to the instructions of Vectastain rabbit ABC Elite kit and ABC kit (Vector Laboratories, Burlingame, CAl, respectively. For staining of frozen sections, antibodies to type III collagen and the Nterminal propeptide of type III pro collagen were diluted to 1:800 and 1:1000, and 10 ~g/ml PBS, and for staining of paraffin sections, to 1:80 and 1:200 and 10-15 ~g/ml PBS, respectively. In some experiments, the blocking serum, and the primary and the secondary antibodies were diluted with PBS containing 20 g, 2 g and 1 g bovine serum albumin/I, respectively. In specificity controls, normal rabbit sera and rabbit immunoglobulin fraction (Dakopatts, Glostrup, Denmark) at appropriate dilutions, and PBS were substituted for the primary antibodies.
Results
The expression patterns of pro-al (I) and Pro-al (III) collagen mRNAs during various stages of mouse and human tooth development were analyzed by in situ hybridization, with use of specific RNA probes. The high degree of conservation of the exon structures and the pattern of exon sizes in genes encoding the fibril-forming collagens (Vuorio and de Crombrugghe, 1990) made it possible to use human probes for the determination of the expression of the two transcripts in mouse tissues. The tissues were considered positive when the relative amounts of silver grains, as visualized by autoradiography, were higher in sections hybridized with the antisense RNA probes than in those hybridized with the corresponding sense probes (serving as negative controls). The expression patterns of pro-al(III) collagen gene transcripts were compared to those of the protein product, as localized by immunohistochemistry. Expression of pro-a 1 (1) and pro-adII1) collagen mRNAs, and (pN) type III collagen during mouse tooth development
In the 13-day-old embryo mandible, the level of proal (III) collagen mRNA expression was moderate in the mesenchyme, except that those mesenchymal cells that
Fig.4. (A, B) The level of pro-ul(III) collagen mRNA expression in the dental pulp (p) of a human premolar tooth germ increases toward the cuspal area, where odontoblasts (0) have started to deposit predentin (pd). The signal showing a moderate intensity is not sharply restricted to the odontoblastic layer. Ameloblasts (a) and stellate reticulum (sr) are negative. (C, D) The signal intensity in the dental follicle (f) is very high, whereas stellate reticulum (sr) is negative. Bright-field (A, C) and dark-field (B, D) images of in situ hybridization patterns with the antisense probe are shown. Bar is 100 flm.
Pro-at(lll) and Pro-aI(I) Collagen mRNAs in Teeth were condensed in the region of the future mandibular bone expressed pro-al (III) transcripts relatively strongly. Weak expression was present in Meckel's cartilage (Fig. 1 A-D). The signal intensity in the oral epithelium did not exceed that in sections hybridized with the sense probe. Immunostaining for the N-terminal propeptide of type III procollagen was moderate in the mesenchyme and strong in the basement membrane zone. Pretreatment of the sections with hyaluronidase facilitated moderate reactivity in the peripheral zone of Meckel's cartilage, while staining in its central region remained weak (Fig. 1 E-G). The use of sagittal sections of the newborn mouse first and second molar regions made it possible to compare the expression of the two mRNAs in dental tissues at different stages of tooth development. During the early bell stage (second molar), the expression of pro-adI) mRNA was weak in the dental papilla while a slightly stronger signal was detected in the cuspal region, where preodontoblasts aligned the odontogenic epithelium. During the late bell stage (first molar), the level of expression was high already in the intercuspal preodontoblasts and even higher in odontoblasts which had become polarized and started to secrete predentin matrix. Hybridization in the dental pulp, oral mesenchyme and follicular tissue surrounding the tooth germs was relatively weak. Expression in osteoblasts and osteocytes, especially in those located on the oral aspect of the developing mandibular bone, was very strong (indicating the direction of bone formation) . The epithelial enamel organ (Fig. 2A, B, E) and the oral epithelium were negative. The expression of pro-ul(III) mRNA in the early bell stage dental papilla was very weak. A clear grain condensate was seen around the tips of the epithelial cervical loop demarcating the dental papilla from the dental follicle, in which the expression was moderate. Deposition of predentin in the cuspal region during the late bell stage involved a gradual increase in the level of pro-al (III) mRNA expression in odontoblasts, but it remained markedly lower than that of pro-at (I) transcripts. The signal intensity in the dental pulp was moderate. Unlike osteoblasts, the mesenchymal cells condensed next to the developing mandibular bone expressed pro-a 1 (III) mRNA strongly (Fig. 2C, D, F). During mineralization of dentin (4-day-old mouse), a moderate signal was detected in odontoblasts. Immunostaining for type III collagen and the N-terminal propeptide of type III procollagen was moderate in the pulp during the late bell stage. Secretory odontoblasts exhibited granular cytoplasmic staining for type III collagen (Fig. 3 A, B), but not for the N-terminal propeptide of type III procollagen. The fairly strong reactivities in the basement membrane zone between preodontoblasts and preameloblasts decreased or disappeared gradually toward the cusp tips, where predentin deposition was most advanced. Pretreatment of the sections with hyaluronidase facilitated weak staining reactions in the last formed (innermost) predentin, and also intensified reactivities in the pulp. Intense staining
509
Fig. 5. In situ hybridization analysis of pro-al(1) collagen mRNA expression in a developing human third molar (about half the root lengths completed). Bright-field (A, C) and dark-field (B, D) images of the hybridization patterns with the antisense probe are shown. (A, B) Odontoblasts (0) on the lateral aspect of the coronal pulp chamber express pro-al(1) transcripts moderately, whereas (C, D) the signal intensity in those situated close to the root apex, where dentin (d) formation is less advanced than coronally, is high. Expression in the pulp (p) (A-D) is weak. Bar is 50 !-lm.
reactions, corresponding to the strong hybridization signal for pro-al(III) mRNA, were seen around the tips of the epithelial cervical loop. The follicular tissue stained strongly, whereas the enamel organ was negative (Fig. 3 C, D). In contrast to the bone matrix, the oral mesenchyme stained intensely. The oral epithelium was negative, except that the superficial keratinizing layer bound type III collagen antibodies non-specifically.
Expression of pro-al (I) and pro-aJCIII) collagen mRNAs, and (pN) type III collagen in human teeth During the stage where mineralization of dentin had started at the tip of the main cusp (second premolar tooth germ), the level of pro-al(III) collagen mRNA expression in the dental pulp increased toward the cuspal area. Moderate
510
P.-L. Lukinmaa et al.
expression was also seen in secretory odontoblasts. Hybridization in the dental follicle was very strong, whereas the epithelial enamel organ was negative (Fig. 4). At a stage where mineralization was in progress throughout the coronal dentin (neonatal tooth), hybridization signal for pro-at (I) mRNA was extremely strong in odontoblasts, and that for pro-at (III) mRNA was moderate. The expression of both transcripts was relatively weak in the pulp, whereas the signal intensities in the gingival connective tissue attached to the cervical part of the tooth crown were high. No expression of either mRNA was observed in the epithelial cervical loop. While granular immunostaining for type III collagen but not for the N-terminal propeptide of type III procollagen was seen in the odontoblastic layer, no reactivities were present in predentin. Both antibodies stained the dental pulp as well as the gingival connective tissue relatively intensely, whereas the cervical loop was negative. No signal for pro-at(III) mRNA was detected in odonto blasts of the two third molars, in which coronal dentin deposition was advanced, but root formation had not begun. Immunostaining for type III collagen and the N-
terminal propeptide of type III procollagen was moderate in the pulp and faint in predentin. In the third molar, in which about half the root lengths had been completed, the very strong expression of pro-at (I) mRNA in odontoblasts elaborating radicular dentin decreased gradually toward the coronal region, where its level, however, remained moderate. The amount of proat (I) transcripts in both the radicular and coronal parts of the pulp was low (Fig. 5). Expression of pro-at (III) mRNA in the odontoblastic layer exhibited a similar gradient with the weak signal in the apical (radicular) odontoblasts disappearing coronally. Hence, the level of pro-at(III) mRNA expression in odontoblasts was clearly lower than that of pro-at (I) mRNA. Hybridization was weak throughout the pulp (Fig. 6 A, B, D, E). Staining intensities for type III collagen and the N-terminal propeptide of type III procollagen in predentin were inversely related to the levels of proat (III) mRNA expression in odontoblasts: whereas no staining was seen in the radicular predentin, reactivities increased gradually toward the coronal predentin. Likewise, reactivities became more intense toward the coronal pulp (Fig. 6 C, F).
Fig. 6. Expression of pro-a! (III) collagen mRNA (A, B, D, E) and the N-terminal propeptide of type III procollagen (C, F) in the developing third molar seen in Fig. 5. Bright-field (A, D) and dark-field (B, E) images of in situ hybridization patterns with the antisense probe are shown. Symbols: p, pulp; 0, odontoblasts; pd, predentin; d, dentin. (A, B) The level of pro-a I (III) mRNA expression in odontoblasts elaborating apical (radicular) predentin slightly exceeds that seen in the pulp. The signal intensity in me developing periodontal ligament (left) is moderate. (C) Immunostaining for the N-terminal propeptide of type III procollagen is relatively weak in the apical pulp and strong in the periodontal ligament (left), whereas no reactivity is present in me apical predentin and dentin. (D, E) Odontoblasts lining the ceiling of the coronal pulp chamber (where dentin formation is advanced) do not express pro-a! (III) transcripts, and the signal intensity in the pulp is low, whereas (F) immunostaining for the N-terminal propeptide of type III procollagen is relatively strong in predentin and the pulp, especially in the sub odontoblastic zone (so). Some reactivity is also detected between odontoblasts. Dentin is negative. Bar is 75 ftm.
Pro-a! (III) and Pro-a! (I) Collagen mRNAs in Teeth
511
d
., :8 .
. p-
... _
- ..
t.' .•
~
.
...... ,
..
4- -..
o Fig. 7. In situ hybridization analysis of pro-al(l) collagen mRNA expression in odontoblasts and the pulp of a fully developed human third molar affected by occlusal caries. Bright-field images illustrating hybridization patterns with the antisense (A, C, D) and sense (B) probes are shown. Symbols; p, pulp; 0, odontoblasts; pd, predentin (C, D); d, dentin. (A, B) The expression is fairly strong in the cuboidal odontoblasts facing the new dentin matrix (nd) deposited in response to the caries lesion, whereas (C) the signal intensity in the normal columnar odontoblasts lining the ceiling of the pulp chamber is weak. (D) Odontoblasts lining the lateral wall of the pulp chamber express pro-al (I) mRNA relatively intensely. Expression in the pulp (A - D) is weak. Bar is 50 !!m.
Fig. 8. (A) Immunostaining for the N-terminal propeptide of type III procollagen is moderate in the pulpal zone of the new dentin matrix (nd) deposited in response to a caries lesion in the tooth shown in Fig. 7. Staining in the pulp (p) is relatively intense. Reactive fiber bundles (arrows) extending from the pulp to the irregular dentin pass the cuboidal odontoblasts (0). A basophilic, intensely mineralized line demarcates the new dentin matrix from normal dentin (top) (pepsin and hyaluronidasepretreated paraffin section). (B) Control section treated with normal rabbit immunoglobulin is negative. (C) Staining for type III collagen is moderate in predentin (pd) and the pulp (p) of a fully developed normal human third molar. The uneven dark color of dentin (top) is due to the mineral phase preserved in the undemineralized frozen section. (D) No staining is seen in a control section incubated with normal rabbit serum. Bar is 50 !!m.
Odontoblasts of a fully developed tooth (affected by dental caries) expressed pro-a! (I) but not pro-a! (III) transcripts. The amount of pro-a! (I) mRNA appeared higher in odontoblasts on the lateral walls of the pulp chamber than in those on its ceiling. The signal intensity in the cuboidal
odontoblasts facing the new dentin matrix deposited in response to the occlusal caries lesion was also high (Fig. 7). Moderate staining for type III collagen and the N-terminal propeptide of type III pro collagen was seen in normal predentin as well as the pulpal zone of the new dentin matrix.
512
P.-L. Lukinmaa et ai.
The pulp stained relatively strongly, and reactive fiber bundles passing the cuboidal odontoblasts extended from the subodontoblastic zone to this irregular matrix (Fig. 8 A, B). Unlike predentin, which stained most consistently in frozen sections (Fig. 8 C, D); the pulp of fully developed teeth stained uniformly with both antibodies, irrespective of the method of tissue preparation (Fig. 8). While reactivities in predentin, when present, frequently increased occlusally, staining for the N-terminal propeptide of type III pro collagen was usually weaker than that for type III collagen and confined to the pulpal zone of predentin. The staining reactions were slightly intensified when the sections were preincubated with hyaluronidase. When normal rabbit serum (contrary to the immunoglobulin fraction) was used as a control, hyaluronidase pretreatment of pepsin -digested sections of formalin-fixed teeth PFA-fixed (mouse) and increased the level of background staining especially in predentin, suggesting that depending on the experimental conditions, predentin has a tendency to bind antibodies/serum components non-specifically. However, since staining in control sections was negligible under experimental conditions/in tissues other than those described above, the staining reactions were considered specific.
Discussion
We have used in situ hybridization analysis to determine the expression of pro-al (III) and pro-al (I) collagen mRNAs in mouse and human dental tissues during tooth development and after its completion. The protein product of the pro-al (III) collagen gene was also localized by immunohistochemistry. It turned out that the expression patterns of the two mRNAs were clearly different, and also that proal (III) transcripts did not show complete codistribution with the protein. Contrary to the current concept, our results suggest that during tooth development, odontoblasts depositing predentin express not only Pro-al (I) but also pro-al (III) collagen gene transcripts. While types I and III collagen codistribute in most connective tissues, the extensive increase in type I collagen synthesis in odontoblasts, once they have become secretory, is generally thought to be accompanied by the cessation of type III collagen production. Interestingly, we found that odontoblasts elaborating predentin matrix expressed proal (III) mRNA, and that the amounts of the transcripts decreased gradually toward the completion of (human) tooth development. Using cDNA from the same source as we, Andujar et ai. (1991) did not detect type III collagen mRNA in odontoblasts of developing mouse teeth, at stages where first coronal odontoblasts had terminally differentiated, or where radicular odontoblasts had become functional. As we did not detect pro-al (III) mRNA in odontoblasts during advanced primary dentinogenesis or after its
completion, the apparent discrepancy between the two studies could be well explained by different developmental stages analyzed. Also, RNA in situ hybridization, used here, may have facilitated the detection of small amounts of proal (III) mRNA. We believe that our results indicate true expression of type III collagen by odontoblasts. First, since the degree of homology between the nucleotide sequences of the proal (III) riboprobe template used, and the corresponding regions of type I pro collagen cDNAs (mRNAs) was low, cross-hybridization is unlikely. Second, the expression patterns of pro-al(I) mRNA and pro-al(I1I) mRNA (and the protein) were clearly different in various cells and tissues such as osteoblasts, follicular tissue and oral mesenchyme. We found that while Pro-al(I1I) mRNA expression in odontoblasts gradually increased with proceeding dentin deposition, and decreased toward the completion of tooth development, immunoreactivity of the protein product in predentin showed a reverse pattern. Also, the expression of pro-al (III) mRNA was weak in the pulp where immunoreactivity of (pN) type III collagen (albeit not necessarily reflecting the amount of the protein), was relatively strong. While the regulation of collagen gene expression is not comprehensively understood (Vuorio and de Crombrugghe, 1990), the correlation between the mRNA expression and the protein production is usually good. A delay in translation could well explain the failure of the protein product to immunostain in predentin simultaneously with pro-al (III) collagen gene transcription in odontoblasts. However, the initial (mouse) predentin stained with both antibodies. While staining of odontoblasts with type III collagen antibodies also supports immediate translation, the failure of these cells to react with antibodies specific for the N-terminal propeptide is difficult to explain. Although adsorption/nonspecific binding of type III collagen antiserum to secretory odontoblasts cannot be definitely excluded, intracytoplasmic conformation of pN type III collagen molecules may alternatively have made the N-terminal propeptide extension inaccessible to antibodies. In terms of the matrix composition, the initial predentin is thought to differ from that elaborated during later stages of dentinogenesis. Also, the conversion of predentin to mineralized dentin involves an increase in the size of collagen fibrils. Type III collagen molecules, many of them retaining the N-terminal propeptide extensions (Fleischmajer et aI., 1981), and thereby believed to restrict the fibril size (Burgeson, 1988), can be located in the same fibrils with type I collagen molecules (Henkel and Glanville, 1982). Hence, the failure to demonstrate (pN) type III collagen immunoreactivity even in predentin during advanced primary dentinogenesis could be due to masking of its antigenic sites by type I collagen molecules in the tight fiber bundles. On the other hand, the appearance of the protein in predentin after the cessation of transcription in odontoblasts toward the completion of primary dentin formation
Pro-a! (III) and Pro-a! (I) Collagen mRNAs in Teeth could be explained not only by the slow rate of matrix deposition but also by the long half-life of collagen, which is thought to be the major factor affecting the accumulation rate of protein. While the long half-life may have contributed to the accumulation of (pN) type III collagen in the dental pulp as well (as suggested by immunostaining), the loose texture and low total collagen content of the pulp may also have rendered the antigenic determinants readily accessible to antibodies. In addition to collagen(s), predentin contains various types of proteoglycan (Takagi et aI., 1990). The enhanced immunoreactivity of (pN) type III collagen in predentin (and the pulp) in sections pretreated with hyaluronidase, known to remove proteoglycans, suggests that masking of the antigenic determinants of this collagen type can at least partially be attributed to proteoglycans. In response to a pathological stimulus such as dental caries, odontoblasts of fully developed teeth can deposit new dentin matrix, which differs morphologically from primary dentin by its more or less irregular structure. While odontoblasts presumed to have deposited this new dentin matrix, and characterized by the intensely stained nuclei and a cuboidal instead of a columnar shape, expressed proat (I) collagen chain mRNA fairly strongly, the transverse plane of cut on normal elongated odontoblasts (leading in fact to an underestimation of the number of grains) makes it difficult to compare the expression levels in morphologically different odontoblasts. On the other hand, the different relative amounts of grains visualized in normal odontoblasts in distinct locations (cut approximately at the same angles in relation to their long axes) could reflect a true difference in the level of pro-at (I) mRNA expression. On the assumption that the correlation between the amount of mRNA and the protein product is good, and also bearing in mind the limited tissue sampling, this would suggest a higher rate of type I collagen synthesis in odontoblasts on the lateral walls of the pulp chamber than in those on its ceiling. The absence of pro-at (III) mRNA in the cuboidal odontoblasts simultaneously with the presence of high amounts of pro-at (I) transcripts (and type III collagen in the new dentin matrix) may suggest that even during reactive matrix deposition, timing of the transcription of the two genes is differential. On the other hand, the presence of fiber bundles extending from the subodontoblastic layer to the new matrix suggests that even pulpal fibroblasts could contribute to its elaboration. The amount of pro-at (III) mRNA in the mesenchymal cells, when condensed in the region of the future mandibular bone was high, but unlike secretory odontoblasts, osteoblasts did not appear to express pro-at (III) transcripts. With regard to pro-a! (III) mRNA expression in cartilage, our results agree with the earlier observations that cells of the articular surfaces of developing human epiphyses express pro-at (III) mRNA (Sandberg et aI., 1989; Treilleux et aI., 1992) and that type III collagen is a
513
minor constituent of cartilage tissue (Keene et aI., 1991; van der Rest, 1991). In summary, our results suggest that secretory odontoblasts express pro-a! (III) collagen mRNA. While its amount increases with the progression of dentinogenesis, and decreases toward the completion of tooth development, these changes are accompanied by a clearly more intense expression of pro-at (I) transcripts. Immunoreactivity of the protein product not only in predentin but also in the pulp in advanced stages of dentinogenesis, simultaneously with no detectable/weak pro-at (III) mRNA expression in odontoblasts and pulpal fibroblasts, respectively, may reflect accumulation of type III collagen with progressing tooth development. This study also emphasizes the usefulness of in situ hybridization and immunohistochemistry as complementary techniques for the elucidation of collagen production in teeth. Cell type- and developmental stage-related differences in the expression of mRNAs for pro-a! (I) and pro-a! (III) collagen chains suggest that the ratio of types I and III collagen influences the structure of dental and periodontal tisues. Acknowledgements We thank Dr. Leena Peltonen for donating the antibodies to type III collagen and Dr. Leila Risteli and Dr. Juha Risteli for the antibodies to the N-terminal propeptide of type III procollagen. We also thank Prof. Jukka Ainamo (Department of Periodontology, Institute of Dentistry, University of Helsinki, Helsinki, Finland) for allowing us to study the human neonatal tooth. The skillful technical assistance of Ms. Marjatta Kivekas, Ms. Sirpa Kyllonen, Ms. Maire Holopainen, Ms. Merja Makinen and Ms. Annikki Siren is gratefully acknowledged. This study was financially supported by the Finnish Dental Society, the Finnish Cultural Foundation, the Academy of Finland, the Sigrid Juselius Foundation and the NIH (grant DE 09399). References Andujar,M.B., Couble,P., Couble,M.-L. and Magloire,H.: Differential expression of type I and type III collagen genes during tooth development. Development 111: 691-698, 1991. Andujar,M.B., Hartmann,D.J., Emonard,H. and Magloire,H.: Distribution and synthesis of type I and type III collagens in developing mouse molar tooth root. Histochemistry 88: 131-140,1988. Becker,J., Schuppan, D., Benzian, H., Bals, T., Hahn, E. G., Cantaluppi, C. and Reichart, P.: Immunohistochemical distribution of collagens types IV, V and VI and of procollagens types I and III in human alveolar bone and dentine. J. Histochem. Cytochem. 34: 1417-1429,1986. Burgeson, R. E.: New collagens, new concepts. Ann. Rev. Cell. BioI. 4: 551-577, 1988. Butler, W. T.: Dentin collagen: chemical structure and role in mineralization. In: Dentin and Dentinogenesis. Vol. II, ed. by Linde, A., CRC Press, Boca Raton, 1984, pp. 37-53. Coumil, I., Leblond, C. P., Pomponio,J., Hand, A. R., Sederlof, L. and Martin, G. R.: Immunohistochemical localization of procollagens. I. Light microscopic distribution of pro collagen I,
514
P.-L. Lukinmaa et al.
III and IV antigenicity in the rat incisor tooth by the indirect peroxidase-antiperoxidase method. ]. Histochem. Cytochem. 27: 1059-1069,1979. Dodd, C. M. and Carmichael, D.].: The collagenous matrix of bovine predentine. Biochim. Biophys. Acta 577: 117-124, 1979. Fleischmajer, R., Timpl, R., Tuderman, L., Raisher, L., Wiestner,M., Perlish,].S. and Graves,P.N.: Ultrastructural identification of extension aminopropeptides of type I and III collagens in human skin. Proc. Natl. Acad. Sci. USA 78: 7360-7364, 1981. Henkel, W. and Glanville, R. W.: Covalent crosslinking between molecules of type I and type III collagen. The involvement of the N-terminal, nonhelical regions of the al(I) and al(lIl) chains in the formation of the intermolecular crosslinks. Eur.]. Biochem. 122: 205-213,1982. Karjalainen, S. and Soderiing, E.: Dentino-pulpal collagen and the incorporation of 3H-proline by sound and carious human teeth in vitro.]. BioI. Buccale 12: 309-316, 1984. Karjalainen, S., Soderiing, E., Pelliniemi, L. and Foidart,].-M.: Immunohistochemical localization of types I and III collagen and fibronectin in the dentine of carious human teeth. Arch. Oral BioI. 31: 801-806, 1986. Keene,D.R., Sakai,L.Y. and Burgeson,R.E.: Human bone contains type III collagen, type VI collagen, and fibrillin: type III collagen is present on specific fibres that may mediate attachment of tendons, ligaments, and periosteum to calcified bone cortex.]. Histochem. Cytochem. 39: 59-69, 1991. Lesot, H., Osman, M. and Ruch,]. V.: Immunofluorescent localization of collagens, fibronectin and laminin during terminal differentiation of odontoblasts. Dev. BioI. 82: 371-381, 1981. Linde,A.: The extracellular matrix of the dental pulp and dentin. ]. Dent. Res. 64: 523-529,1985. Linde, A.: Dentin matrix proteins: composition and possible functions in calcification. Anat. Rec. 224: 154-166,1989. Lukinmaa,P.-L. and Waltimo,].: Immunohistochemicallocalization of types I, V and VI collagen in human permanent teeth and periodontalligament.]. Dent. Res. 71: 391-397, 1992. Lukinmaa, P.-L., Vaahtokari, A., Vainio, S. and Thesleff, I.: Expression of type I collagen pro-a2 chain mRNA in adult human permanent teeth as revealed by in situ hybridization. ]. Dent. Res. 71: 36-42, 1992. Munksgaard, E. C. and Moe, D.: Type of collagen in an extract of odontoblasts and dentine from developing bovine teeth. Arch. Oral Bioi. 25: 485-489, 1980. Munksgaard, E. c., Rhodes, M., Mayne, R. and Butler, W. T.: Collagen synthesis and secretion by rat incisor odontoblasts in organ culture. Eur.]. Biochem. 82: 609-617, 1978. Nagata, K., Huang, Y. H., Ohsaki, Y., Kukita, T., Nakata, M. and Kurisu, K.: Demonstration of Type III Collagen in the Dentin of Mice. Matrix 12: 448-455, 1992. Niemela,O., Risteli, L., Parkkinen,j. and Risteli,].: Purification and characterization of the N-terminal propeptide of human typeIIIprocollagen. Biochem.]. 32: 145-150,1985. Peltonen, L., Myllyla, R., Tolonen, U. and Myllyla, V. V.: Changes in collagen metabolism in diseased muscle. II. Immunohistochemical studies. Arch. Neurol. 39: 756-759, 1982. Ruch,]. V.: Determinisms of odontogenesis. Cell. BioI. Rev. RBC 14: 1-112, 1987. Sandberg, M., Autio-Harmainen, H. and Vuorio, E.: Localization of the expression of types I, III and IV collagen, TGF-131 and cfos genes in developing human calvarial bone. Dev. BioI. 130: 324-334,1988.
Sandberg,M., Makela,j.K., Muitimaki,P., Vuorio, T. and Vuorio,E.: Construction of human pro-a 1(III) collagen eDNA clone and localization of type III collagen expression in human fetal tissues. Matrix 9: 82-91, 1989. Sandberg, M. and Vuorio, E.: Localization of types I, II and III collagen mRNAs in developing human skeletal tissues by in situ hybridization.]. Cell. Bioi. 104: 1077-1084,1987. Scott, P. G. and Veis, A.: The cyanogen peptides of bovine soluble and insoluble collagens. II. Tissue-specific cross-linked peptides of insoluble skin and dentin collagen. Connect. Tissue Res. 4: 117-129,1976. Shuttleworth,C.A., Ward,j.L. and Hirschmann,P.N.: The presence of type III collagen in developing tooth. Biochim. Biophys. Acta 535: 348-355, 1978. Sodek,]. and Mandell, S.M.: Collagen metabolism in rat incisor predentine in vivo: Synthesis and maturation of type!, al(I) trimer and typeV collagens. Biochemistry 21: 2011-2015, 1982. Takagi,M., Hishikawa,H., Hosokawa, Y., Kagami,A. and Rahemtulla, F.: Immunohistochemical localization of glycosaminoglycans and proteoglycans in predentin and dentin of rat incisors.]. Histochem. Cytochem. 38: 319-324, 1990. Takita,K., Ohsaki, Y., Nakata,M. and Kurisu,K.: Immunofluorescence localization of type I and type III collagen and fibronectin in mouse dental tissues in late development and during molar eruption. Arch. Oral BioI. 32: 273-279, 1987. Thesleff, I., Partanen, A.-M. and Vainio, S.: Epithelial-mesenchymal interactions in tooth morphogenesis: The roles of extracellular matrix, growth factors, and cell surface receptors. ]. Craniofac. Genet. Dev. Bioi. 11: 229-237, 1991. Thesleff, I., Stenman, S., Vaheri, A. and Timpl, R.: Changes in the matrix proteins, fibronectin and collagen, during differentiation of mouse tooth germ. Dev. Bioi. 70: 116-126,1979. Treilleux, I., Mallein-Gerin, F., Ie Guellec, D. and Herbage, D.: Localization of the expression of type I, II, III collagen, and aggrecan core protein genes in developing human articular cartilage. Matrix 12: 221-232, 1992. Tsuzaki,M., Yamauchi,M. and Mechanic,G.L.: Bovine dental pulp collagens: Characterization of types III and V collagen. Arch. Oral Bioi. 35: 195-200, 1990. Tung,P.S., Domenicucci,C., Wasi,S. and Sodek,j.: Specific immunohistochemical localization of osteonectin and collagen types I and III in fetal and adult porcine dental tissues. ]. Histochem. Cytochem. 33: 531-540,1985. van Amerongen,j.P., Lemmens,I.G. and Tonino,G.].M.: The concentration, extractability and characterization of collagen in human dental pulp. Arch. Oral Bioi. 28: 339-345, 1983. van der Rest, M.: The collagens of bone. In: Bone Matrix and Bone Specific Products, Vol. 3, ed. by Hall, B. K., CRC Press, Boca Raton, 1991, pp. 187-237. Volpin,D. and Veis, A.: Cyanogen bromide peptides from insoluble skin and dentin collagens. Biochemistry 12: 1452-1464, 1973. . Vuorio, E. and de Crombrugghe, B.: The family of collagen genes. Ann. Rev. Biochem. 59: 837-872, 1990. Vuorio, T., Makela,j.K., Kahari, V.-M. and Vuorio, E.: Coordinated regulation of type I and type III collagen production and mRNA levels of pro al(l) and pro a2(I) collagen in cultured morphea fibroblasts. Arch. Dermatol. Res. 279: 154-160, 1987. Wilkinson, D. G. and Green,j.: In situ hybridization and the threedimensional reconstruction of serial sections. In: Postimplantation Mammalian Embryos: A Practical Approach, ed. by
Pro-a1 (III) and Pro-a1 (I) Collagen mRNAs in Teeth Copp, A.J. and Cockroft, D. L., Oxford University Press, London, 1990, pp. 155 -171. Wright,G.M. and Leblond,C.P.: An attempt at quantitation of procollagens I and III and of collagen IV in tooth sections by exposure to the corresponding antibodies followed by 1251_pro_
tein A and radioautography.
1355 -1358,1980.
J.
515
Histochem. Cytochem. 28:
Dr. Pirjo-Liisa Lukinmaa, Department of Oral Pathology, Institute of Dentistry, P. O. Box 41 (Mannerheimintie 172), SF00014 University of Helsinki, Finland.
Transient Expression of Type III Collagen by Odontoblasts: Developmental Changes in the Distribution of Pro-u1 (III) and Pro-u1 (I) Collagen mRNAs in Dental Tissues PIRJO-LiISA LUKINMAA1, ANNE VAAHTOKARI 2 , SEPPO VAINI0 2 , MINNA SANDBERG 3 , JANNA WALTIM02 and IRMA THESLEFF 2 1 2
3
Department of Oral Pathology, Institute of Dentistry , University of Helsinki, Helsinki, Department of Pedodontics and Orthodontics, Institute of Dentistry , University of Helsinki, Helsinki and Department of Medical Biochemistry, University of Turku, Turku, Finland.
Abstract
The expression of Pro-al (III) and pro-al (I) collagen mRNAs in mouse and human dental tissues during tooth development and after its completion was analyzed by in situ hybridization, with use of [35 S]-labeled RNA probes. The expression of pro-a I (III) mRNA was also compared to that of the protein product, as localized by immunostaining with polyclonal antibodies to type III collagen and the N-terminal propeptide of type III procollagen. Contrary to many previous reports, our results suggest that odontoblasts express type III collagen. While proal (III) transcripts were less intensely expressed in odontoblasts than pro-a I (I) transcripts, the amounts of both mRNAs increased in odontoblasts with progressing dentin formation, and decreased toward its completion. In contrast to Pro-al (III) mRNA, pro-al (I) mRNA was still detectable in odontoblasts of fully developed teeth. Type III collagen immunoreactivity was observed in the early predentin, and again in predentin toward the completion of dentinogenesis, when mRNA was no longer detected. Also in the pulp, the protein product, unlike pro-al(III) mRNA, was relatively strongly expressed. Hence, these immunostaining patterns were inversely related to the expression of Pro-a1 (III) mRNA, suggesting accumulation of the protein. The mesenchymal cells, when condensed in the region of the future mandibular bone, expressed proal (III) mRNA intensely, whereas osteoblasts expressed pro-aj (I) but not pro-aj (III) transcripts strongly. Cell type- and developmental stage-related differences in the expression of the two mRNAs suggest that type Iltype III collagen ratio influences the structure of dental tissues. Key words: in situ hybridization, immunohistochemistry, pro-a I (III) and Pro-al (I) collagen mRNAs, (pN) type III collagen, teeth. Introduction
The main structural component of the organic matrices of predentin and dentin of various species is type I collagen (for reviews, see Butler, 1984; Linde, 1989). While type III collagen remains a major constituent of the dental pulp even after the completion of tooth development (Shuttleworth et al., 1978; van Amerongen et al., 1983; Linde,
1985; Tsuzaki et al., 1990), it is generally thought that odontoblasts do not express type III collagen at the mRNA (Andujar et al., 1991) or protein level (Lesot et al., 1981; Ruch, 1987; Andujar et al., 1988). Although early biochemical studies have suggested the presence of type III collagen in dentin (Volpin and Veis, 1973), in a number of later biochemical (Scott and Veis, 1976; Munksgaard et al., 1978; Dodd and Carmichael, 1979; Munksgaard and Moe,
504
P.-L. Lukinmaa et al.
1980; Sodek and Mandell, 1982; Karjalainen and Soderling, 1984) and immunohistochemical (Cournil et ai., 1979; Wright and Leblond, 1980; Takita et ai., 1987) studies on teeth of different species it has not been possible to demonstrate that odontoblasts synthesize type III collagen or that it is present in predentin. On the other hand, some immunohistochemical studies suggest that terminal differentiation of odontoblasts might not be accompanied by the cessation of type III collagen production. For example, type III collagen immunoreactivity has been detected in odontoblasts of fetal porcine and adult human teeth (Tung et ai., 1985; Karjalainen et ai., 1986). Also the first formed predentin of developing mouse teeth (Thesleff, et ai., 1979, 1991), dentin of young mice (Nagata et ai., 1992) as well as predentin of fully developed human teeth (Becker et ai., 1986; Karjalainen et ai., 1986) have been found to immunostain for type III (pro )collagen. Osteoblasts are believed to represent another major cell lineage that does not express type III collagen, as demonstrated even at the mRNA level (Sandberg and Vuorio, 1987; Sandberg et ai., 1988; Andujar et ai., 1991). However, recent immunoelectron microscopic data show that not only developing but also adult human cortical bone contains type III collagen specifically retained in discrete fiber bundles (Keene et ai., 1991). While the genes coding for the pro-a chains of types I and III collagen are differentially expressed in distinct tissues and at various stages of development (Burgeson, 1988; Sandberg et ai., 1989; Andujar et ai., 1991), the production of these proteins usually correlates well with the expression of the corresponding mRNAs (Sandberg and Vuorio, 1987; Sandberg et ai., 1989). With regard to teeth, once tooth development has been completed, the rate of dentin deposition is markedly decreased and unlike bone, dentin is not remodeled. Furthermore, as the half-life of collagen is remarkably long, the presence of the protein product in predentin does not necessarily imply simultaneous transcription in odontoblasts. We have previously shown that after the completion of human tooth development, odontoblasts continue to express mRNA for pro-a2(I) collagen chain (Lukinmaa et ai., 1992). The aim of this study was to analyze the expression of the genes encoding the pro-al chains of types I and III collagen in dental tissues. For this purpose, the expression of the two mRNAs was determined in mouse and human dental tissues during tooth development and after its completion, with radiolabeled RNA probes. The expression of pro-al (III) collagen transcripts was also compared to that of the protein product, as localized by immunohistochemical staining with antibodies to type III collagen and the N-terminal propeptide of type III pro collagen.
Materials and Methods Mouse Teeth
The regions of the mandibular molar tooth germs were dissected from 13-day-old embryonic (vaginal plug = day 0), newborn, and 4-day-old postnatal mice (CBAxC57BL). As determined morphologically, the developmental stages of the first and second molars covered the bud stage (not analyzed), the early (odontoblasts not polarized) and late (deposition of predentin initiated in the cuspal region) bell stages, and a stage where dentin was being mineralized throughout the tooth crown. Human Teeth
The germ of the mandibular left second premolar tooth (mineralization of dentin started at the tip of the main cusp) was accidentally removed from a four and a half-year-old girl, in association with the therapeutic removal of the deciduous second molar. A neonatal mandibular incisor (mineralization of dentin in progress throughout the tooth crown) was obtained from a newborn baby. Two mandibular third molars, in which coronal dentin formation was advanced but root development had not begun, were removed from a 14-year-old girl. One developing (about half the root lengths completed) and several fully developed third molars, one of them (to be analyzed by in situ hybridization) affected by occlusal caries, were from young adults. These teeth were also removed for therapeutic reasons. Preparation of the tissues for in situ hybridization
The mouse molar regions were fixed overnight with 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS) at 4°C, dehydrated through a rising ethanol series and xylene, embedded in paraffin, sectioned at 71-lm thickness on silane-coated slides, and air-dried at 37°C. The human teeth with the attached soft tissues were fixed with 10% neutral buffered formalin or 4% PFA for 2-5 days. To facilitate the penetration of the fixative to the pulp chamber of the fully developed tooth, about one third of the root lengths was dissected away immediately after extraction. The teeth were demineralized with ethylenediaminetetra-acetic acid (0.33 MIl) for 4-12 weeks (depending on the stage of development) at 4°C, rinsed overnight with running tap water, and processed into paraffin sections as described above. The slides were stored in tight boxes at 4 °C until used. For a detailed histological examination, some sections, corresponding to those used in the hybridization experiments, were stained with hematoxylin and eosin.
Pro-a}(III) and Pro-al(I) Collagen mRNAs in Teeth
Preparation of the probes A 372-bp Pst I-Pvu II fragment from clone pHCALl, specific for human pro-a} (I) collagen mRNA (Vuorio et ai., 1987), was sub cloned into pGEMl plasmid (Promega Biotec, Madison, WI). The plasmid was linearized with Pst I or EcoR I restriction endonuclease (Boehringer Mannheim, Germany), and eSS]-UTP-labeled single-stranded antisense and sense pro-a} (I) RNAs were synthesized in vitro (Wilkinson and Green, 1990), with T7 or SP6 RNA polymerase (Promega Biotec), respectively. A 295-bp Pst I-Pst I fragment from clone pHFS3, specific for human pro-a} (III) collagen mRNA (Sandberg et al., 1989), was also subcloned into pGEM1 plasmid, which was linearized with Hind III or
505
EcoR I restriction endonuclease (Boehringer Mannheim). For synthesis of the antisense and sense pro-a} (III) RNAs, T7 or SP6 RNA polymerase was used. RNAs were shortened to an average length of 100 bases by limited alkaline hydrolysis. For isolation ofprobes of the desired length, gelfiltration chromatography was performed on a Sephadex G-50 (Pharmacia, Uppsala, Sweden) column. After precipitation with ethanol, the probes were dissolved in the hybridization buffer containing 60% formamide, and diluted to about 5.5 X 104 dpm[ll-1. Homologies between the nucleotide sequences of the pro-a} (I) collagen probe and the corresponding regions of pro-a2(I) and pro-a}(III) mRNAs were 64% and 59%,
Fig. I. Expression of pro-al(III) collagen mRNA (A-D) and the N-terminal propeptide of type III procollagen (E-G) in transverse sections through a 13-day-old mouse embryo mandible. In situ hybridization patterns with the antisense (A, B) and sense (negative control) (C, D) probes, and bright-field (A, C) and dark-field (B, D) images are shown. The level of pro-adIII) mRNA expression is moderate in the mesenchyme (m). Those mesenchymal cells that are condensed in the region of the future mandibular bone (arrows) express pro-al(III) transcripts relatively strongly. A weak signal is also detected in Meckel's cartilage (c). (E) Immunostaining for the Nterminal propeptide of type III pro collagen is moderate in the mesenchyme (m), whereas Meckel's cartilage (c) is negative, unless the section is pretreated with hyaluronidase, which facilitates reactivity especially in its peripheral zone, as seen in (F). (G) No staining is seen in a hyaluronidase-pretreated control section incubated with normal rabbit immunoglobulin. Bar is 50 ftm.
506
P.-L. Lukinmaa et al.
respectively. The nucleotide sequence of the pro-at (III) probe showed only 51 % identity with the corresponding regions of the human pro-at (I) and pro-a2(I) collagen mRNAs. The probes did not contain the sequence of the 87
nucleotides in the middle of the carboxy-terminal propeptide domain, which is virtually identical in mRNAs for fibril-forming collagens (Sandberg and Vuorio, 1987).
Fig. 2. Pro-at (I) (A, B, E) and pro-at (Ill) (C, D, F) collagen mRNA expression in sagittal sections through the mandibular molar region of a newborn mouse. Bright-field (A, C, E, F) and dark-field (B, D) images of in situ hybridization patterns with the antisense probes are shown. The first (right; A-D) and second (left) molar tooth germs correspond to the late and early bell stages, respectively. Symbols: 0, odontoblasts; p, dental pulp/papilla (A-D); f, dental follicle (A-E); a, ameloblasts; sr, stellate reticulum; m, oral mesenchyme; b, mandibular bone. (A, B) Expression of pro-at (I) mRNA in the cuspal area (arrows) of the second molar germ is slightly stronger than elsewhere in the dental papilla. Polarized odontoblasts (first molar) express pro-at(l) mRNA very strongly. Hybridization is relatively weak in the dental pulp, follicular tissue and oral mesenchyme, whereas the signal intensity in the developing mandibular bone is fairly high. Ameloblasts and stellate reticulum are negative. (C, D) Virtually no expression of pro-at (III) mRNA is seen in the dental papilla of the second molar. With deposition of predentin (first molar), the level of expression increases in odontoblasts, and the signal in the pulp is also more intense than in the dental papilla of the second molar. Grain condensations (arrowheads) are seen around the epithelial cervical loop tips. Hybridization is moderate in the follicular tissue and relatively strong in the oral mesenchyme. Ameloblasts and stellate reticulum are negative. Higher magnifications (E, F) show that osteoblasts (arrows), especially those on the oral aspect of the developing mandibular bone, express (E) pro-at(I) but not (F) pro-at (III) collagen transcripts strongly. Expression of pro-at (III) mRNA in fibroblasts between the bony trabeculae is intense. Also note that the signal intensity for pro-at (III) transcripts is lower in odontoblasts of the first molar cusp tip and higher in the oral mesenchyme than that for pro-at (I) mRNA. Oral epithelium (e) is negative (F). Bar is 150 ~m in A-D and 50 ~m in E and F.
Pro-at (III) and Pro-at (I) Collagen mRNAs in Teeth
In situ hybridization In situ hybridization was performed as described by Wilkinson and Green (1990). Tissue sections were rehydrated, treated with proteinase K and acetylated. Before hybridization, the probes were heated at 80 DC for 2 min. The sections were hybridized overnight at 50-52 DC, in a tight box moistured with the hybridization buffer containing 60% formamide. The hybridized sections were washed under highly stringent conditions, including a wash at 65 DC (30 min), a treatment with ribonuclease A (Boehringer Mannheim, Germany), and another 30 min wash at 65 DC. The slides were air-dried, dipped in autoradiographic emulsion (Kodak NTB-2; dilution 1:1), again dried in the presence of silica gel, and exposed at 4 DC for 10-15 days. After development and fixation, the sections were shortly counterstained with hematoxylin, and mounted with DePex (BDH, Poole, England). The slides were examined under a Leitz Orthoplan microscope and selectively photographed using Ilford Pan F film.
507
Preparation of the specimens for immunohistochemistry Paraffin sections of mouse and human tissues were prepared as described above. Mouse tissues were also fixed with 94% cold ethanol and human teeth with ethanol at about 22 DC. To enhance the reactivities, some rehydrated sections were pretreated with pepsin (Merck, Darmstadt, Germany; 0.4% in 0.1 M HCI, 37"C, 40 min) and/or ovine testicular hyaluronidase (Boehringer Mannheim, Germany; 3500 and 7000U/ml PBS, pH6.0, 37"C, 30 min). Some human teeth were frozen with liquid nitrogen and processed into sections as earlier described (Lukinmaa and Waltimo, 1992).
Antibodies Specific rabbit anticalf skin type III collagen antisera, cross-reactive with tissues of other species (Peltonen et al., 1982), were generously donated by Dr. Leena Peltonen (National Public Health Institute, Helsinki, Finland). Immunoaffinity-purified rabbit antibodies specific for the N-terminal propeptide of human type III pro collagen
Fig. 3. (A) Immunostaining for type III collagen in the late bell stage first molar tooth germ of a newborn mouse is moderate in the pulp (p) and cytoplasmic reactivity is also seen in secretory odontoblasts (0). Predentin (pd) in the enzymatically untreated section is virtually negative. The dental follicle (f) stains very intensely, whereas the ameloblastic layer (a) is negative. (B) Control section treated with normal rabbit serum is negative. (C) Antibodies to the N-terminal propeptide of type III procollagen stain the pulp (p) of the first molar moderately, whereas the odontoblastic layer (0) is negative. The strong reactivity in the dental basement membrane zone (arrows) disappears toward predentin (arrowheads) in the cuspal area, where (D) hyaluronidase pretreatment facilitates weak staining in the innermost predentin (arrowheads) and also intensifies reactivity in the pulp (p) and basement membrane. Staining in the dental follicle (f) is intense, whereas the ameloblastic layer (a) and stellate reticulum (sr) are negative. Note the strong reactivity around the epithelial cervical loop (cl) (C, D). Bar is 30 ~m in A and B, and 100 ~m in C and D.
508
P.-L. Lukinmaa et al.
(Niemela et aI., 1985) were a kind gift from Drs. Leila and Juha Risteli (Department of Medical Biochemistry, University of Oulu, Oulu, Finland). Immunohistochemical staining
Frozen and paraffin sections were stained with the immunoperoxidase method as earlier described (Lukinmaa and Waltimo, 1992), according to the instructions of Vectastain rabbit ABC Elite kit and ABC kit (Vector Laboratories, Burlingame, CAl, respectively. For staining of frozen sections, antibodies to type III collagen and the Nterminal propeptide of type III pro collagen were diluted to 1:800 and 1:1000, and 10 ~g/ml PBS, and for staining of paraffin sections, to 1:80 and 1:200 and 10-15 ~g/ml PBS, respectively. In some experiments, the blocking serum, and the primary and the secondary antibodies were diluted with PBS containing 20 g, 2 g and 1 g bovine serum albumin/I, respectively. In specificity controls, normal rabbit sera and rabbit immunoglobulin fraction (Dakopatts, Glostrup, Denmark) at appropriate dilutions, and PBS were substituted for the primary antibodies.
Results
The expression patterns of pro-al (I) and Pro-al (III) collagen mRNAs during various stages of mouse and human tooth development were analyzed by in situ hybridization, with use of specific RNA probes. The high degree of conservation of the exon structures and the pattern of exon sizes in genes encoding the fibril-forming collagens (Vuorio and de Crombrugghe, 1990) made it possible to use human probes for the determination of the expression of the two transcripts in mouse tissues. The tissues were considered positive when the relative amounts of silver grains, as visualized by autoradiography, were higher in sections hybridized with the antisense RNA probes than in those hybridized with the corresponding sense probes (serving as negative controls). The expression patterns of pro-al(III) collagen gene transcripts were compared to those of the protein product, as localized by immunohistochemistry. Expression of pro-a 1 (1) and pro-adII1) collagen mRNAs, and (pN) type III collagen during mouse tooth development
In the 13-day-old embryo mandible, the level of proal (III) collagen mRNA expression was moderate in the mesenchyme, except that those mesenchymal cells that
Fig.4. (A, B) The level of pro-ul(III) collagen mRNA expression in the dental pulp (p) of a human premolar tooth germ increases toward the cuspal area, where odontoblasts (0) have started to deposit predentin (pd). The signal showing a moderate intensity is not sharply restricted to the odontoblastic layer. Ameloblasts (a) and stellate reticulum (sr) are negative. (C, D) The signal intensity in the dental follicle (f) is very high, whereas stellate reticulum (sr) is negative. Bright-field (A, C) and dark-field (B, D) images of in situ hybridization patterns with the antisense probe are shown. Bar is 100 flm.
Pro-at(lll) and Pro-aI(I) Collagen mRNAs in Teeth were condensed in the region of the future mandibular bone expressed pro-al (III) transcripts relatively strongly. Weak expression was present in Meckel's cartilage (Fig. 1 A-D). The signal intensity in the oral epithelium did not exceed that in sections hybridized with the sense probe. Immunostaining for the N-terminal propeptide of type III procollagen was moderate in the mesenchyme and strong in the basement membrane zone. Pretreatment of the sections with hyaluronidase facilitated moderate reactivity in the peripheral zone of Meckel's cartilage, while staining in its central region remained weak (Fig. 1 E-G). The use of sagittal sections of the newborn mouse first and second molar regions made it possible to compare the expression of the two mRNAs in dental tissues at different stages of tooth development. During the early bell stage (second molar), the expression of pro-adI) mRNA was weak in the dental papilla while a slightly stronger signal was detected in the cuspal region, where preodontoblasts aligned the odontogenic epithelium. During the late bell stage (first molar), the level of expression was high already in the intercuspal preodontoblasts and even higher in odontoblasts which had become polarized and started to secrete predentin matrix. Hybridization in the dental pulp, oral mesenchyme and follicular tissue surrounding the tooth germs was relatively weak. Expression in osteoblasts and osteocytes, especially in those located on the oral aspect of the developing mandibular bone, was very strong (indicating the direction of bone formation) . The epithelial enamel organ (Fig. 2A, B, E) and the oral epithelium were negative. The expression of pro-ul(III) mRNA in the early bell stage dental papilla was very weak. A clear grain condensate was seen around the tips of the epithelial cervical loop demarcating the dental papilla from the dental follicle, in which the expression was moderate. Deposition of predentin in the cuspal region during the late bell stage involved a gradual increase in the level of pro-al (III) mRNA expression in odontoblasts, but it remained markedly lower than that of pro-at (I) transcripts. The signal intensity in the dental pulp was moderate. Unlike osteoblasts, the mesenchymal cells condensed next to the developing mandibular bone expressed pro-a 1 (III) mRNA strongly (Fig. 2C, D, F). During mineralization of dentin (4-day-old mouse), a moderate signal was detected in odontoblasts. Immunostaining for type III collagen and the N-terminal propeptide of type III procollagen was moderate in the pulp during the late bell stage. Secretory odontoblasts exhibited granular cytoplasmic staining for type III collagen (Fig. 3 A, B), but not for the N-terminal propeptide of type III procollagen. The fairly strong reactivities in the basement membrane zone between preodontoblasts and preameloblasts decreased or disappeared gradually toward the cusp tips, where predentin deposition was most advanced. Pretreatment of the sections with hyaluronidase facilitated weak staining reactions in the last formed (innermost) predentin, and also intensified reactivities in the pulp. Intense staining
509
Fig. 5. In situ hybridization analysis of pro-al(1) collagen mRNA expression in a developing human third molar (about half the root lengths completed). Bright-field (A, C) and dark-field (B, D) images of the hybridization patterns with the antisense probe are shown. (A, B) Odontoblasts (0) on the lateral aspect of the coronal pulp chamber express pro-al(1) transcripts moderately, whereas (C, D) the signal intensity in those situated close to the root apex, where dentin (d) formation is less advanced than coronally, is high. Expression in the pulp (p) (A-D) is weak. Bar is 50 !-lm.
reactions, corresponding to the strong hybridization signal for pro-al(III) mRNA, were seen around the tips of the epithelial cervical loop. The follicular tissue stained strongly, whereas the enamel organ was negative (Fig. 3 C, D). In contrast to the bone matrix, the oral mesenchyme stained intensely. The oral epithelium was negative, except that the superficial keratinizing layer bound type III collagen antibodies non-specifically.
Expression of pro-al (I) and pro-aJCIII) collagen mRNAs, and (pN) type III collagen in human teeth During the stage where mineralization of dentin had started at the tip of the main cusp (second premolar tooth germ), the level of pro-al(III) collagen mRNA expression in the dental pulp increased toward the cuspal area. Moderate
510
P.-L. Lukinmaa et al.
expression was also seen in secretory odontoblasts. Hybridization in the dental follicle was very strong, whereas the epithelial enamel organ was negative (Fig. 4). At a stage where mineralization was in progress throughout the coronal dentin (neonatal tooth), hybridization signal for pro-at (I) mRNA was extremely strong in odontoblasts, and that for pro-at (III) mRNA was moderate. The expression of both transcripts was relatively weak in the pulp, whereas the signal intensities in the gingival connective tissue attached to the cervical part of the tooth crown were high. No expression of either mRNA was observed in the epithelial cervical loop. While granular immunostaining for type III collagen but not for the N-terminal propeptide of type III procollagen was seen in the odontoblastic layer, no reactivities were present in predentin. Both antibodies stained the dental pulp as well as the gingival connective tissue relatively intensely, whereas the cervical loop was negative. No signal for pro-at(III) mRNA was detected in odonto blasts of the two third molars, in which coronal dentin deposition was advanced, but root formation had not begun. Immunostaining for type III collagen and the N-
terminal propeptide of type III procollagen was moderate in the pulp and faint in predentin. In the third molar, in which about half the root lengths had been completed, the very strong expression of pro-at (I) mRNA in odontoblasts elaborating radicular dentin decreased gradually toward the coronal region, where its level, however, remained moderate. The amount of proat (I) transcripts in both the radicular and coronal parts of the pulp was low (Fig. 5). Expression of pro-at (III) mRNA in the odontoblastic layer exhibited a similar gradient with the weak signal in the apical (radicular) odontoblasts disappearing coronally. Hence, the level of pro-at(III) mRNA expression in odontoblasts was clearly lower than that of pro-at (I) mRNA. Hybridization was weak throughout the pulp (Fig. 6 A, B, D, E). Staining intensities for type III collagen and the N-terminal propeptide of type III procollagen in predentin were inversely related to the levels of proat (III) mRNA expression in odontoblasts: whereas no staining was seen in the radicular predentin, reactivities increased gradually toward the coronal predentin. Likewise, reactivities became more intense toward the coronal pulp (Fig. 6 C, F).
Fig. 6. Expression of pro-a! (III) collagen mRNA (A, B, D, E) and the N-terminal propeptide of type III procollagen (C, F) in the developing third molar seen in Fig. 5. Bright-field (A, D) and dark-field (B, E) images of in situ hybridization patterns with the antisense probe are shown. Symbols: p, pulp; 0, odontoblasts; pd, predentin; d, dentin. (A, B) The level of pro-a I (III) mRNA expression in odontoblasts elaborating apical (radicular) predentin slightly exceeds that seen in the pulp. The signal intensity in me developing periodontal ligament (left) is moderate. (C) Immunostaining for the N-terminal propeptide of type III procollagen is relatively weak in the apical pulp and strong in the periodontal ligament (left), whereas no reactivity is present in me apical predentin and dentin. (D, E) Odontoblasts lining the ceiling of the coronal pulp chamber (where dentin formation is advanced) do not express pro-a! (III) transcripts, and the signal intensity in the pulp is low, whereas (F) immunostaining for the N-terminal propeptide of type III procollagen is relatively strong in predentin and the pulp, especially in the sub odontoblastic zone (so). Some reactivity is also detected between odontoblasts. Dentin is negative. Bar is 75 ftm.
Pro-a! (III) and Pro-a! (I) Collagen mRNAs in Teeth
511
d
., :8 .
. p-
... _
- ..
t.' .•
~
.
...... ,
..
4- -..
o Fig. 7. In situ hybridization analysis of pro-al(l) collagen mRNA expression in odontoblasts and the pulp of a fully developed human third molar affected by occlusal caries. Bright-field images illustrating hybridization patterns with the antisense (A, C, D) and sense (B) probes are shown. Symbols; p, pulp; 0, odontoblasts; pd, predentin (C, D); d, dentin. (A, B) The expression is fairly strong in the cuboidal odontoblasts facing the new dentin matrix (nd) deposited in response to the caries lesion, whereas (C) the signal intensity in the normal columnar odontoblasts lining the ceiling of the pulp chamber is weak. (D) Odontoblasts lining the lateral wall of the pulp chamber express pro-al (I) mRNA relatively intensely. Expression in the pulp (A - D) is weak. Bar is 50 !!m.
Fig. 8. (A) Immunostaining for the N-terminal propeptide of type III procollagen is moderate in the pulpal zone of the new dentin matrix (nd) deposited in response to a caries lesion in the tooth shown in Fig. 7. Staining in the pulp (p) is relatively intense. Reactive fiber bundles (arrows) extending from the pulp to the irregular dentin pass the cuboidal odontoblasts (0). A basophilic, intensely mineralized line demarcates the new dentin matrix from normal dentin (top) (pepsin and hyaluronidasepretreated paraffin section). (B) Control section treated with normal rabbit immunoglobulin is negative. (C) Staining for type III collagen is moderate in predentin (pd) and the pulp (p) of a fully developed normal human third molar. The uneven dark color of dentin (top) is due to the mineral phase preserved in the undemineralized frozen section. (D) No staining is seen in a control section incubated with normal rabbit serum. Bar is 50 !!m.
Odontoblasts of a fully developed tooth (affected by dental caries) expressed pro-a! (I) but not pro-a! (III) transcripts. The amount of pro-a! (I) mRNA appeared higher in odontoblasts on the lateral walls of the pulp chamber than in those on its ceiling. The signal intensity in the cuboidal
odontoblasts facing the new dentin matrix deposited in response to the occlusal caries lesion was also high (Fig. 7). Moderate staining for type III collagen and the N-terminal propeptide of type III pro collagen was seen in normal predentin as well as the pulpal zone of the new dentin matrix.
512
P.-L. Lukinmaa et ai.
The pulp stained relatively strongly, and reactive fiber bundles passing the cuboidal odontoblasts extended from the subodontoblastic zone to this irregular matrix (Fig. 8 A, B). Unlike predentin, which stained most consistently in frozen sections (Fig. 8 C, D); the pulp of fully developed teeth stained uniformly with both antibodies, irrespective of the method of tissue preparation (Fig. 8). While reactivities in predentin, when present, frequently increased occlusally, staining for the N-terminal propeptide of type III pro collagen was usually weaker than that for type III collagen and confined to the pulpal zone of predentin. The staining reactions were slightly intensified when the sections were preincubated with hyaluronidase. When normal rabbit serum (contrary to the immunoglobulin fraction) was used as a control, hyaluronidase pretreatment of pepsin -digested sections of formalin-fixed teeth PFA-fixed (mouse) and increased the level of background staining especially in predentin, suggesting that depending on the experimental conditions, predentin has a tendency to bind antibodies/serum components non-specifically. However, since staining in control sections was negligible under experimental conditions/in tissues other than those described above, the staining reactions were considered specific.
Discussion
We have used in situ hybridization analysis to determine the expression of pro-al (III) and pro-al (I) collagen mRNAs in mouse and human dental tissues during tooth development and after its completion. The protein product of the pro-al (III) collagen gene was also localized by immunohistochemistry. It turned out that the expression patterns of the two mRNAs were clearly different, and also that proal (III) transcripts did not show complete codistribution with the protein. Contrary to the current concept, our results suggest that during tooth development, odontoblasts depositing predentin express not only Pro-al (I) but also pro-al (III) collagen gene transcripts. While types I and III collagen codistribute in most connective tissues, the extensive increase in type I collagen synthesis in odontoblasts, once they have become secretory, is generally thought to be accompanied by the cessation of type III collagen production. Interestingly, we found that odontoblasts elaborating predentin matrix expressed proal (III) mRNA, and that the amounts of the transcripts decreased gradually toward the completion of (human) tooth development. Using cDNA from the same source as we, Andujar et ai. (1991) did not detect type III collagen mRNA in odontoblasts of developing mouse teeth, at stages where first coronal odontoblasts had terminally differentiated, or where radicular odontoblasts had become functional. As we did not detect pro-al (III) mRNA in odontoblasts during advanced primary dentinogenesis or after its
completion, the apparent discrepancy between the two studies could be well explained by different developmental stages analyzed. Also, RNA in situ hybridization, used here, may have facilitated the detection of small amounts of proal (III) mRNA. We believe that our results indicate true expression of type III collagen by odontoblasts. First, since the degree of homology between the nucleotide sequences of the proal (III) riboprobe template used, and the corresponding regions of type I pro collagen cDNAs (mRNAs) was low, cross-hybridization is unlikely. Second, the expression patterns of pro-al(I) mRNA and pro-al(I1I) mRNA (and the protein) were clearly different in various cells and tissues such as osteoblasts, follicular tissue and oral mesenchyme. We found that while Pro-al(I1I) mRNA expression in odontoblasts gradually increased with proceeding dentin deposition, and decreased toward the completion of tooth development, immunoreactivity of the protein product in predentin showed a reverse pattern. Also, the expression of pro-al (III) mRNA was weak in the pulp where immunoreactivity of (pN) type III collagen (albeit not necessarily reflecting the amount of the protein), was relatively strong. While the regulation of collagen gene expression is not comprehensively understood (Vuorio and de Crombrugghe, 1990), the correlation between the mRNA expression and the protein production is usually good. A delay in translation could well explain the failure of the protein product to immunostain in predentin simultaneously with pro-al (III) collagen gene transcription in odontoblasts. However, the initial (mouse) predentin stained with both antibodies. While staining of odontoblasts with type III collagen antibodies also supports immediate translation, the failure of these cells to react with antibodies specific for the N-terminal propeptide is difficult to explain. Although adsorption/nonspecific binding of type III collagen antiserum to secretory odontoblasts cannot be definitely excluded, intracytoplasmic conformation of pN type III collagen molecules may alternatively have made the N-terminal propeptide extension inaccessible to antibodies. In terms of the matrix composition, the initial predentin is thought to differ from that elaborated during later stages of dentinogenesis. Also, the conversion of predentin to mineralized dentin involves an increase in the size of collagen fibrils. Type III collagen molecules, many of them retaining the N-terminal propeptide extensions (Fleischmajer et aI., 1981), and thereby believed to restrict the fibril size (Burgeson, 1988), can be located in the same fibrils with type I collagen molecules (Henkel and Glanville, 1982). Hence, the failure to demonstrate (pN) type III collagen immunoreactivity even in predentin during advanced primary dentinogenesis could be due to masking of its antigenic sites by type I collagen molecules in the tight fiber bundles. On the other hand, the appearance of the protein in predentin after the cessation of transcription in odontoblasts toward the completion of primary dentin formation
Pro-a! (III) and Pro-a! (I) Collagen mRNAs in Teeth could be explained not only by the slow rate of matrix deposition but also by the long half-life of collagen, which is thought to be the major factor affecting the accumulation rate of protein. While the long half-life may have contributed to the accumulation of (pN) type III collagen in the dental pulp as well (as suggested by immunostaining), the loose texture and low total collagen content of the pulp may also have rendered the antigenic determinants readily accessible to antibodies. In addition to collagen(s), predentin contains various types of proteoglycan (Takagi et aI., 1990). The enhanced immunoreactivity of (pN) type III collagen in predentin (and the pulp) in sections pretreated with hyaluronidase, known to remove proteoglycans, suggests that masking of the antigenic determinants of this collagen type can at least partially be attributed to proteoglycans. In response to a pathological stimulus such as dental caries, odontoblasts of fully developed teeth can deposit new dentin matrix, which differs morphologically from primary dentin by its more or less irregular structure. While odontoblasts presumed to have deposited this new dentin matrix, and characterized by the intensely stained nuclei and a cuboidal instead of a columnar shape, expressed proat (I) collagen chain mRNA fairly strongly, the transverse plane of cut on normal elongated odontoblasts (leading in fact to an underestimation of the number of grains) makes it difficult to compare the expression levels in morphologically different odontoblasts. On the other hand, the different relative amounts of grains visualized in normal odontoblasts in distinct locations (cut approximately at the same angles in relation to their long axes) could reflect a true difference in the level of pro-at (I) mRNA expression. On the assumption that the correlation between the amount of mRNA and the protein product is good, and also bearing in mind the limited tissue sampling, this would suggest a higher rate of type I collagen synthesis in odontoblasts on the lateral walls of the pulp chamber than in those on its ceiling. The absence of pro-at (III) mRNA in the cuboidal odontoblasts simultaneously with the presence of high amounts of pro-at (I) transcripts (and type III collagen in the new dentin matrix) may suggest that even during reactive matrix deposition, timing of the transcription of the two genes is differential. On the other hand, the presence of fiber bundles extending from the subodontoblastic layer to the new matrix suggests that even pulpal fibroblasts could contribute to its elaboration. The amount of pro-at (III) mRNA in the mesenchymal cells, when condensed in the region of the future mandibular bone was high, but unlike secretory odontoblasts, osteoblasts did not appear to express pro-at (III) transcripts. With regard to pro-a! (III) mRNA expression in cartilage, our results agree with the earlier observations that cells of the articular surfaces of developing human epiphyses express pro-at (III) mRNA (Sandberg et aI., 1989; Treilleux et aI., 1992) and that type III collagen is a
513
minor constituent of cartilage tissue (Keene et aI., 1991; van der Rest, 1991). In summary, our results suggest that secretory odontoblasts express pro-a! (III) collagen mRNA. While its amount increases with the progression of dentinogenesis, and decreases toward the completion of tooth development, these changes are accompanied by a clearly more intense expression of pro-at (I) transcripts. Immunoreactivity of the protein product not only in predentin but also in the pulp in advanced stages of dentinogenesis, simultaneously with no detectable/weak pro-at (III) mRNA expression in odontoblasts and pulpal fibroblasts, respectively, may reflect accumulation of type III collagen with progressing tooth development. This study also emphasizes the usefulness of in situ hybridization and immunohistochemistry as complementary techniques for the elucidation of collagen production in teeth. Cell type- and developmental stage-related differences in the expression of mRNAs for pro-a! (I) and pro-a! (III) collagen chains suggest that the ratio of types I and III collagen influences the structure of dental and periodontal tisues. Acknowledgements We thank Dr. Leena Peltonen for donating the antibodies to type III collagen and Dr. Leila Risteli and Dr. Juha Risteli for the antibodies to the N-terminal propeptide of type III procollagen. We also thank Prof. Jukka Ainamo (Department of Periodontology, Institute of Dentistry, University of Helsinki, Helsinki, Finland) for allowing us to study the human neonatal tooth. The skillful technical assistance of Ms. Marjatta Kivekas, Ms. Sirpa Kyllonen, Ms. Maire Holopainen, Ms. Merja Makinen and Ms. Annikki Siren is gratefully acknowledged. This study was financially supported by the Finnish Dental Society, the Finnish Cultural Foundation, the Academy of Finland, the Sigrid Juselius Foundation and the NIH (grant DE 09399). References Andujar,M.B., Couble,P., Couble,M.-L. and Magloire,H.: Differential expression of type I and type III collagen genes during tooth development. Development 111: 691-698, 1991. Andujar,M.B., Hartmann,D.J., Emonard,H. and Magloire,H.: Distribution and synthesis of type I and type III collagens in developing mouse molar tooth root. Histochemistry 88: 131-140,1988. Becker,J., Schuppan, D., Benzian, H., Bals, T., Hahn, E. G., Cantaluppi, C. and Reichart, P.: Immunohistochemical distribution of collagens types IV, V and VI and of procollagens types I and III in human alveolar bone and dentine. J. Histochem. Cytochem. 34: 1417-1429,1986. Burgeson, R. E.: New collagens, new concepts. Ann. Rev. Cell. BioI. 4: 551-577, 1988. Butler, W. T.: Dentin collagen: chemical structure and role in mineralization. In: Dentin and Dentinogenesis. Vol. II, ed. by Linde, A., CRC Press, Boca Raton, 1984, pp. 37-53. Coumil, I., Leblond, C. P., Pomponio,J., Hand, A. R., Sederlof, L. and Martin, G. R.: Immunohistochemical localization of procollagens. I. Light microscopic distribution of pro collagen I,
514
P.-L. Lukinmaa et al.
III and IV antigenicity in the rat incisor tooth by the indirect peroxidase-antiperoxidase method. ]. Histochem. Cytochem. 27: 1059-1069,1979. Dodd, C. M. and Carmichael, D.].: The collagenous matrix of bovine predentine. Biochim. Biophys. Acta 577: 117-124, 1979. Fleischmajer, R., Timpl, R., Tuderman, L., Raisher, L., Wiestner,M., Perlish,].S. and Graves,P.N.: Ultrastructural identification of extension aminopropeptides of type I and III collagens in human skin. Proc. Natl. Acad. Sci. USA 78: 7360-7364, 1981. Henkel, W. and Glanville, R. W.: Covalent crosslinking between molecules of type I and type III collagen. The involvement of the N-terminal, nonhelical regions of the al(I) and al(lIl) chains in the formation of the intermolecular crosslinks. Eur.]. Biochem. 122: 205-213,1982. Karjalainen, S. and Soderiing, E.: Dentino-pulpal collagen and the incorporation of 3H-proline by sound and carious human teeth in vitro.]. BioI. Buccale 12: 309-316, 1984. Karjalainen, S., Soderiing, E., Pelliniemi, L. and Foidart,].-M.: Immunohistochemical localization of types I and III collagen and fibronectin in the dentine of carious human teeth. Arch. Oral BioI. 31: 801-806, 1986. Keene,D.R., Sakai,L.Y. and Burgeson,R.E.: Human bone contains type III collagen, type VI collagen, and fibrillin: type III collagen is present on specific fibres that may mediate attachment of tendons, ligaments, and periosteum to calcified bone cortex.]. Histochem. Cytochem. 39: 59-69, 1991. Lesot, H., Osman, M. and Ruch,]. V.: Immunofluorescent localization of collagens, fibronectin and laminin during terminal differentiation of odontoblasts. Dev. BioI. 82: 371-381, 1981. Linde,A.: The extracellular matrix of the dental pulp and dentin. ]. Dent. Res. 64: 523-529,1985. Linde, A.: Dentin matrix proteins: composition and possible functions in calcification. Anat. Rec. 224: 154-166,1989. Lukinmaa,P.-L. and Waltimo,].: Immunohistochemicallocalization of types I, V and VI collagen in human permanent teeth and periodontalligament.]. Dent. Res. 71: 391-397, 1992. Lukinmaa, P.-L., Vaahtokari, A., Vainio, S. and Thesleff, I.: Expression of type I collagen pro-a2 chain mRNA in adult human permanent teeth as revealed by in situ hybridization. ]. Dent. Res. 71: 36-42, 1992. Munksgaard, E. C. and Moe, D.: Type of collagen in an extract of odontoblasts and dentine from developing bovine teeth. Arch. Oral Bioi. 25: 485-489, 1980. Munksgaard, E. c., Rhodes, M., Mayne, R. and Butler, W. T.: Collagen synthesis and secretion by rat incisor odontoblasts in organ culture. Eur.]. Biochem. 82: 609-617, 1978. Nagata, K., Huang, Y. H., Ohsaki, Y., Kukita, T., Nakata, M. and Kurisu, K.: Demonstration of Type III Collagen in the Dentin of Mice. Matrix 12: 448-455, 1992. Niemela,O., Risteli, L., Parkkinen,j. and Risteli,].: Purification and characterization of the N-terminal propeptide of human typeIIIprocollagen. Biochem.]. 32: 145-150,1985. Peltonen, L., Myllyla, R., Tolonen, U. and Myllyla, V. V.: Changes in collagen metabolism in diseased muscle. II. Immunohistochemical studies. Arch. Neurol. 39: 756-759, 1982. Ruch,]. V.: Determinisms of odontogenesis. Cell. BioI. Rev. RBC 14: 1-112, 1987. Sandberg, M., Autio-Harmainen, H. and Vuorio, E.: Localization of the expression of types I, III and IV collagen, TGF-131 and cfos genes in developing human calvarial bone. Dev. BioI. 130: 324-334,1988.
Sandberg,M., Makela,j.K., Muitimaki,P., Vuorio, T. and Vuorio,E.: Construction of human pro-a 1(III) collagen eDNA clone and localization of type III collagen expression in human fetal tissues. Matrix 9: 82-91, 1989. Sandberg, M. and Vuorio, E.: Localization of types I, II and III collagen mRNAs in developing human skeletal tissues by in situ hybridization.]. Cell. Bioi. 104: 1077-1084,1987. Scott, P. G. and Veis, A.: The cyanogen peptides of bovine soluble and insoluble collagens. II. Tissue-specific cross-linked peptides of insoluble skin and dentin collagen. Connect. Tissue Res. 4: 117-129,1976. Shuttleworth,C.A., Ward,j.L. and Hirschmann,P.N.: The presence of type III collagen in developing tooth. Biochim. Biophys. Acta 535: 348-355, 1978. Sodek,]. and Mandell, S.M.: Collagen metabolism in rat incisor predentine in vivo: Synthesis and maturation of type!, al(I) trimer and typeV collagens. Biochemistry 21: 2011-2015, 1982. Takagi,M., Hishikawa,H., Hosokawa, Y., Kagami,A. and Rahemtulla, F.: Immunohistochemical localization of glycosaminoglycans and proteoglycans in predentin and dentin of rat incisors.]. Histochem. Cytochem. 38: 319-324, 1990. Takita,K., Ohsaki, Y., Nakata,M. and Kurisu,K.: Immunofluorescence localization of type I and type III collagen and fibronectin in mouse dental tissues in late development and during molar eruption. Arch. Oral BioI. 32: 273-279, 1987. Thesleff, I., Partanen, A.-M. and Vainio, S.: Epithelial-mesenchymal interactions in tooth morphogenesis: The roles of extracellular matrix, growth factors, and cell surface receptors. ]. Craniofac. Genet. Dev. Bioi. 11: 229-237, 1991. Thesleff, I., Stenman, S., Vaheri, A. and Timpl, R.: Changes in the matrix proteins, fibronectin and collagen, during differentiation of mouse tooth germ. Dev. Bioi. 70: 116-126,1979. Treilleux, I., Mallein-Gerin, F., Ie Guellec, D. and Herbage, D.: Localization of the expression of type I, II, III collagen, and aggrecan core protein genes in developing human articular cartilage. Matrix 12: 221-232, 1992. Tsuzaki,M., Yamauchi,M. and Mechanic,G.L.: Bovine dental pulp collagens: Characterization of types III and V collagen. Arch. Oral Bioi. 35: 195-200, 1990. Tung,P.S., Domenicucci,C., Wasi,S. and Sodek,j.: Specific immunohistochemical localization of osteonectin and collagen types I and III in fetal and adult porcine dental tissues. ]. Histochem. Cytochem. 33: 531-540,1985. van Amerongen,j.P., Lemmens,I.G. and Tonino,G.].M.: The concentration, extractability and characterization of collagen in human dental pulp. Arch. Oral Bioi. 28: 339-345, 1983. van der Rest, M.: The collagens of bone. In: Bone Matrix and Bone Specific Products, Vol. 3, ed. by Hall, B. K., CRC Press, Boca Raton, 1991, pp. 187-237. Volpin,D. and Veis, A.: Cyanogen bromide peptides from insoluble skin and dentin collagens. Biochemistry 12: 1452-1464, 1973. . Vuorio, E. and de Crombrugghe, B.: The family of collagen genes. Ann. Rev. Biochem. 59: 837-872, 1990. Vuorio, T., Makela,j.K., Kahari, V.-M. and Vuorio, E.: Coordinated regulation of type I and type III collagen production and mRNA levels of pro al(l) and pro a2(I) collagen in cultured morphea fibroblasts. Arch. Dermatol. Res. 279: 154-160, 1987. Wilkinson, D. G. and Green,j.: In situ hybridization and the threedimensional reconstruction of serial sections. In: Postimplantation Mammalian Embryos: A Practical Approach, ed. by
Pro-a1 (III) and Pro-a1 (I) Collagen mRNAs in Teeth Copp, A.J. and Cockroft, D. L., Oxford University Press, London, 1990, pp. 155 -171. Wright,G.M. and Leblond,C.P.: An attempt at quantitation of procollagens I and III and of collagen IV in tooth sections by exposure to the corresponding antibodies followed by 1251_pro_
tein A and radioautography.
1355 -1358,1980.
J.
515
Histochem. Cytochem. 28:
Dr. Pirjo-Liisa Lukinmaa, Department of Oral Pathology, Institute of Dentistry, P. O. Box 41 (Mannerheimintie 172), SF00014 University of Helsinki, Finland.